An ERDAS image processing method for retrieving LST and describing urban heat evolution: a case study in the Pearl River Delta Region in South China

A spatial-temporal model with Model Maker tool is designed to retrieve Land Surface Temperature (LST) and to describe the changes of urban heat island, as well as urban development. Spectral Radiance, Brightness Temperature, NDVI, and Emissivity are first calculated from TM and ETM+, which are then used to compute LST by using Qin et al.’s mono-window algorithm. The LST is classified based on normalized statistical method, and the normalized heat images are computed between different times. Therefore, the urban heat changes can be shown in the map clearly and directly through an urban heat conversion matrix. Such a model has been applied in this study to obtain the urban heat conversion matrix of South China from 1990 to 2000. The results indicate that the LST increased areas mainly locate along the major roads in the eastern bank of the Pearl River, which is a result of speedy urban expansion and need to be noticed in the future.